Enhanced Single-Shot Beam Training for True-Time-Delay Hybrid Beamforming Architecture

Massive arrays are used in mmWave channels to compensate for the severe propagation path loss. One of the common challenges is the beam steering and selection. In this work, we use true-time-delay to minimize the beam training overhead associated with large number of antennas in mmWave. We propose an angle-of-arrival estimation algorithm that leverages the degrees of freedom introduced by hybrid beamforming system and thus, allows for single-shot, i.e., single-pilot, estimation process with good performance. The proposed algorithm improves the 5
^th
percentile beamforming gain compared to the existing single-shot algorithms. Our simulations show that the proposed algorithm enhances the receiver’s sensitivity by 3 dB when using hybrid arrays. We show that at relatively fair SNR, the number of analog-to-digital converters at the receiver side can be decreased without significant performance loss, thus saving power and area.